Message ID | 20170421074157.19159-1-hdegoede@redhat.com (mailing list archive) |
---|---|
State | Not Applicable, archived |
Headers | show |
On Fri, 21 Apr 2017 09:41:57 +0200, Hans de Goede wrote: > > Some peripherals on Baytrail and Cherrytrail platforms signal PME to the > PMC to wakeup the system. When this happens software needs to clear the > PME_B0_STS bit in the GPE0a_STS register to avoid an IRQ storm on IRQ 9. > > This is modelled in ACPI through the INT0002 ACPI device. > > This commit adds a driver which will bind to that device, call the > ACPI event handler for the wakeup and clear the interrupt source > avoiding the irq storm. > > Signed-off-by: Hans de Goede <hdegoede@redhat.com> This patch requires your previous "always present" patchset, right? Some nitpicking: > --- /dev/null > +++ b/drivers/platform/x86/intel_int0002.c ... > +struct int0002_data { > + struct spinlock lock; > + struct device *dev; > + const struct x86_cpu_id *cpu_id; > + acpi_handle handle; > + char ev_name[5]; Are 5 bytes enough? I see the code below: > + snprintf(data->ev_name, sizeof(data->ev_name), "_%c%02X", > + res->data.gpio.triggering ? 'E' : 'L', > + res->data.gpio.pin_table[0]); So it counts 6 including NUL. > +static int int0002_probe(struct platform_device *pdev) > +{ > + struct acpi_buffer buf = { ACPI_ALLOCATE_BUFFER, NULL }; > + struct device *dev = &pdev->dev; > + struct int0002_data *data; > + struct acpi_resource *res; > + acpi_status status; > + acpi_handle hdl; > + int irq, ret; > + > + data = devm_kzalloc(dev, sizeof(*data), GFP_KERNEL); > + if (!data) { > + dev_err(dev, "can't allocate memory for int0002\n"); The error message is mostly superfluous. > +static int int0002_runtime_suspend(struct device *dev) > +{ > + return 0; > +} > + > +static int int0002_runtime_resume(struct device *dev) > +{ > + return 0; > +} > + > +static const struct dev_pm_ops int0002_pm_ops = { > + .runtime_suspend = int0002_runtime_suspend, > + .runtime_resume = int0002_runtime_resume, > +}; Do we need these runtime PM? If not, we can remove the header inclusion, too. Other than that, I confirmed to work on my ASUS E200H. Tested-by: Takashi Iwai <tiwai@suse.de> thanks, Takashi -- To unsubscribe from this list: send the line "unsubscribe linux-acpi" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Fri, Apr 21, 2017 at 11:17 AM, Takashi Iwai <tiwai@suse.de> wrote: > On Fri, 21 Apr 2017 09:41:57 +0200, > Hans de Goede wrote: >> >> Some peripherals on Baytrail and Cherrytrail platforms signal PME to the >> PMC to wakeup the system. When this happens software needs to clear the >> PME_B0_STS bit in the GPE0a_STS register to avoid an IRQ storm on IRQ 9. >> >> This is modelled in ACPI through the INT0002 ACPI device. >> >> This commit adds a driver which will bind to that device, call the >> ACPI event handler for the wakeup and clear the interrupt source >> avoiding the irq storm. >> + char ev_name[5]; > > Are 5 bytes enough? I see the code below: > >> + snprintf(data->ev_name, sizeof(data->ev_name), "_%c%02X", >> + res->data.gpio.triggering ? 'E' : 'L', >> + res->data.gpio.pin_table[0]); > > So it counts 6 including NUL. How? 4 + NUL = 5. OTOH it looks like code duplication with existing drivers (GPIO ACPI library IIRC) which might make sense to make generic. >> + data = devm_kzalloc(dev, sizeof(*data), GFP_KERNEL); >> + if (!data) { >> + dev_err(dev, "can't allocate memory for int0002\n"); > > The error message is mostly superfluous. > >> +static int int0002_runtime_suspend(struct device *dev) >> +{ >> + return 0; >> +} >> + >> +static int int0002_runtime_resume(struct device *dev) >> +{ >> + return 0; >> +} >> + >> +static const struct dev_pm_ops int0002_pm_ops = { >> + .runtime_suspend = int0002_runtime_suspend, >> + .runtime_resume = int0002_runtime_resume, >> +}; > > Do we need these runtime PM? If not, we can remove the header > inclusion, too. Yeah, and it needs attention when built with !CONFIG_PM.
On Fri, 21 Apr 2017 10:27:32 +0200, Andy Shevchenko wrote: > > On Fri, Apr 21, 2017 at 11:17 AM, Takashi Iwai <tiwai@suse.de> wrote: > > On Fri, 21 Apr 2017 09:41:57 +0200, > > Hans de Goede wrote: > >> > >> Some peripherals on Baytrail and Cherrytrail platforms signal PME to the > >> PMC to wakeup the system. When this happens software needs to clear the > >> PME_B0_STS bit in the GPE0a_STS register to avoid an IRQ storm on IRQ 9. > >> > >> This is modelled in ACPI through the INT0002 ACPI device. > >> > >> This commit adds a driver which will bind to that device, call the > >> ACPI event handler for the wakeup and clear the interrupt source > >> avoiding the irq storm. > > >> + char ev_name[5]; > > > > Are 5 bytes enough? I see the code below: > > > >> + snprintf(data->ev_name, sizeof(data->ev_name), "_%c%02X", > >> + res->data.gpio.triggering ? 'E' : 'L', > >> + res->data.gpio.pin_table[0]); > > > > So it counts 6 including NUL. > > How? 4 + NUL = 5. Well, "_E00X" is 5 letters, IIUC. > OTOH it looks like code duplication with existing drivers (GPIO ACPI > library IIRC) which might make sense to make generic. > > >> + data = devm_kzalloc(dev, sizeof(*data), GFP_KERNEL); > >> + if (!data) { > >> + dev_err(dev, "can't allocate memory for int0002\n"); > > > > The error message is mostly superfluous. > > > >> +static int int0002_runtime_suspend(struct device *dev) > >> +{ > >> + return 0; > >> +} > >> + > >> +static int int0002_runtime_resume(struct device *dev) > >> +{ > >> + return 0; > >> +} > >> + > >> +static const struct dev_pm_ops int0002_pm_ops = { > >> + .runtime_suspend = int0002_runtime_suspend, > >> + .runtime_resume = int0002_runtime_resume, > >> +}; > > > > Do we need these runtime PM? If not, we can remove the header > > inclusion, too. > > Yeah, and it needs attention when built with !CONFIG_PM. Practically seen, we may build this only with CONFIG_PM, too. The virtual GPIO thing happens only when the machine gets resumed. Takashi -- To unsubscribe from this list: send the line "unsubscribe linux-acpi" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Fri, Apr 21, 2017 at 11:38 AM, Takashi Iwai <tiwai@suse.de> wrote: > On Fri, 21 Apr 2017 10:27:32 +0200, > Andy Shevchenko wrote: >> On Fri, Apr 21, 2017 at 11:17 AM, Takashi Iwai <tiwai@suse.de> wrote: >> > On Fri, 21 Apr 2017 09:41:57 +0200, >> > Hans de Goede wrote: >> >> + char ev_name[5]; >> > >> > Are 5 bytes enough? I see the code below: >> > >> >> + snprintf(data->ev_name, sizeof(data->ev_name), "_%c%02X", >> >> + res->data.gpio.triggering ? 'E' : 'L', >> >> + res->data.gpio.pin_table[0]); >> > >> > So it counts 6 including NUL. >> >> How? 4 + NUL = 5. > > Well, "_E00X" is 5 letters, IIUC. Well, today is Friday, yes. But %02X means two capitalized hex characters. >> >> +static int int0002_runtime_resume(struct device *dev) >> >> +{ >> >> + return 0; >> >> +} >> >> + >> >> +static const struct dev_pm_ops int0002_pm_ops = { >> >> + .runtime_suspend = int0002_runtime_suspend, >> >> + .runtime_resume = int0002_runtime_resume, >> >> +}; >> > >> > Do we need these runtime PM? If not, we can remove the header >> > inclusion, too. >> >> Yeah, and it needs attention when built with !CONFIG_PM. > > Practically seen, we may build this only with CONFIG_PM, too. > The virtual GPIO thing happens only when the machine gets resumed. Perhaps depend on PM then?
On Fri, 21 Apr 2017 11:02:20 +0200, Andy Shevchenko wrote: > > On Fri, Apr 21, 2017 at 11:38 AM, Takashi Iwai <tiwai@suse.de> wrote: > > On Fri, 21 Apr 2017 10:27:32 +0200, > > Andy Shevchenko wrote: > >> On Fri, Apr 21, 2017 at 11:17 AM, Takashi Iwai <tiwai@suse.de> wrote: > >> > On Fri, 21 Apr 2017 09:41:57 +0200, > >> > Hans de Goede wrote: > > >> >> + char ev_name[5]; > >> > > >> > Are 5 bytes enough? I see the code below: > >> > > >> >> + snprintf(data->ev_name, sizeof(data->ev_name), "_%c%02X", > >> >> + res->data.gpio.triggering ? 'E' : 'L', > >> >> + res->data.gpio.pin_table[0]); > >> > > >> > So it counts 6 including NUL. > >> > >> How? 4 + NUL = 5. > > > > Well, "_E00X" is 5 letters, IIUC. > > Well, today is Friday, yes. > But %02X means two capitalized hex characters. Bah, of course, Friday morning, and I have to take a coffee. > >> >> +static int int0002_runtime_resume(struct device *dev) > >> >> +{ > >> >> + return 0; > >> >> +} > >> >> + > >> >> +static const struct dev_pm_ops int0002_pm_ops = { > >> >> + .runtime_suspend = int0002_runtime_suspend, > >> >> + .runtime_resume = int0002_runtime_resume, > >> >> +}; > >> > > >> > Do we need these runtime PM? If not, we can remove the header > >> > inclusion, too. > >> > >> Yeah, and it needs attention when built with !CONFIG_PM. > > > > Practically seen, we may build this only with CONFIG_PM, too. > > The virtual GPIO thing happens only when the machine gets resumed. > > Perhaps depend on PM then? Yes, it may make our lives a bit easier. We stumbled on this driver just because of the hang at resume, and who wants Cherrytrail device without PM? thanks, Takashi -- To unsubscribe from this list: send the line "unsubscribe linux-acpi" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Hi, On 21-04-17 10:17, Takashi Iwai wrote: > On Fri, 21 Apr 2017 09:41:57 +0200, > Hans de Goede wrote: >> >> Some peripherals on Baytrail and Cherrytrail platforms signal PME to the >> PMC to wakeup the system. When this happens software needs to clear the >> PME_B0_STS bit in the GPE0a_STS register to avoid an IRQ storm on IRQ 9. >> >> This is modelled in ACPI through the INT0002 ACPI device. >> >> This commit adds a driver which will bind to that device, call the >> ACPI event handler for the wakeup and clear the interrupt source >> avoiding the irq storm. >> >> Signed-off-by: Hans de Goede <hdegoede@redhat.com> > > This patch requires your previous "always present" patchset, right? On boards where the _STA for INT0002 returns 0, that patchset is needed for the driver to bind to the device, so yes it has a functional requirement on some (most it seems) board, but only at run-time, it can be merged independently. > Some nitpicking: > >> --- /dev/null >> +++ b/drivers/platform/x86/intel_int0002.c > ... >> +struct int0002_data { >> + struct spinlock lock; >> + struct device *dev; >> + const struct x86_cpu_id *cpu_id; >> + acpi_handle handle; >> + char ev_name[5]; > > Are 5 bytes enough? I see the code below: > >> + snprintf(data->ev_name, sizeof(data->ev_name), "_%c%02X", >> + res->data.gpio.triggering ? 'E' : 'L', >> + res->data.gpio.pin_table[0]); > > So it counts 6 including NUL. The name is e.g. _E02 or _L02 so 5 bytes including the NUL is enough :) > >> +static int int0002_probe(struct platform_device *pdev) >> +{ >> + struct acpi_buffer buf = { ACPI_ALLOCATE_BUFFER, NULL }; >> + struct device *dev = &pdev->dev; >> + struct int0002_data *data; >> + struct acpi_resource *res; >> + acpi_status status; >> + acpi_handle hdl; >> + int irq, ret; >> + >> + data = devm_kzalloc(dev, sizeof(*data), GFP_KERNEL); >> + if (!data) { >> + dev_err(dev, "can't allocate memory for int0002\n"); > > The error message is mostly superfluous. Agreed, will fix for v2. >> +static int int0002_runtime_suspend(struct device *dev) >> +{ >> + return 0; >> +} >> + >> +static int int0002_runtime_resume(struct device *dev) >> +{ >> + return 0; >> +} >> + >> +static const struct dev_pm_ops int0002_pm_ops = { >> + .runtime_suspend = int0002_runtime_suspend, >> + .runtime_resume = int0002_runtime_resume, >> +}; > > Do we need these runtime PM? If not, we can remove the header > inclusion, too. We enable runtime pm for this device, at which point these are mandatory. > Other than that, I confirmed to work on my ASUS E200H. > Tested-by: Takashi Iwai <tiwai@suse.de> Regards, Hans -- To unsubscribe from this list: send the line "unsubscribe linux-acpi" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Fri, Apr 21, 2017 at 09:41:57AM +0200, Hans de Goede wrote: > Some peripherals on Baytrail and Cherrytrail platforms signal PME to the > PMC to wakeup the system. When this happens software needs to clear the > PME_B0_STS bit in the GPE0a_STS register to avoid an IRQ storm on IRQ 9. > > This is modelled in ACPI through the INT0002 ACPI device. > > This commit adds a driver which will bind to that device, call the > ACPI event handler for the wakeup and clear the interrupt source > avoiding the irq storm. > > Signed-off-by: Hans de Goede <hdegoede@redhat.com> > --- > drivers/platform/x86/Kconfig | 10 ++ > drivers/platform/x86/Makefile | 1 + > drivers/platform/x86/intel_int0002.c | 232 +++++++++++++++++++++++++++++++++++ > 3 files changed, 243 insertions(+) > create mode 100644 drivers/platform/x86/intel_int0002.c > [...snip] > --- /dev/null > +++ b/drivers/platform/x86/intel_int0002.c > @@ -0,0 +1,232 @@ > +/* > + * Intel INT0002 "Virtual GPIO" driver > + * > + * Copyright (C) 2017 Hans de Goede <hdegoede@redhat.com> > + * > + * Loosely based on android x86 kernel code which is: > + * > + * Copyright (c) 2014, Intel Corporation. > + * > + * Author: Dyut Kumar Sil <dyut.k.sil@intel.com> > + * > + * This program is free software; you can redistribute it and/or modify > + * it under the terms of the GNU General Public License version 2 as > + * published by the Free Software Foundation. > + * > + * Some peripherals on Baytrail and Cherrytrail platforms signal PME to the > + * PMC to wakeup the system. When this happens software needs to clear the > + * PME_B0_STS bit in the GPE0a_STS register to avoid an IRQ storm on IRQ 9. > + * > + * This is modelled in ACPI through the INT0002 ACPI device, which is > + * called a "Virtual GPIO controller" in ACPI because it defines the event > + * handler to call when the PME triggers through _AEI and _L02 / _E02 > + * methods as would be done for a real GPIO interrupt. > + * > + * This driver will bind to the INT0002 device, call the ACPI event handler > + * for the wakeup and clear the interrupt source avoiding the irq storm. > + */ > + > +#include <asm/cpu_device_id.h> > +#include <asm/intel-family.h> > +#include <asm/io.h> > +#include <linux/acpi.h> > +#include <linux/interrupt.h> > +#include <linux/kernel.h> > +#include <linux/module.h> > +#include <linux/platform_device.h> > +#include <linux/pm_runtime.h> > +#include <linux/slab.h> > + > +#define GPE0A_PME_STS_BIT 0x2000 > +#define GPE0A_PME_EN_BIT 0x2000 Maybe it's minor. Could you please consider to change the name to GPE0A_PME_B0_STS_BIT and GPE0A_PME_B0_EN_BIT? Because it's easy to confuse with the PME_STS/PME_EN bit (0x800, bit 11) Thanks a lot! Joey Lee -- To unsubscribe from this list: send the line "unsubscribe linux-acpi" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Hi, On 21-04-17 12:11, Hans de Goede wrote: > Hi, > > On 21-04-17 10:17, Takashi Iwai wrote: >> On Fri, 21 Apr 2017 09:41:57 +0200, >> Hans de Goede wrote: >>> >>> Some peripherals on Baytrail and Cherrytrail platforms signal PME to the >>> PMC to wakeup the system. When this happens software needs to clear the >>> PME_B0_STS bit in the GPE0a_STS register to avoid an IRQ storm on IRQ 9. >>> >>> This is modelled in ACPI through the INT0002 ACPI device. >>> >>> This commit adds a driver which will bind to that device, call the >>> ACPI event handler for the wakeup and clear the interrupt source >>> avoiding the irq storm. >>> >>> Signed-off-by: Hans de Goede <hdegoede@redhat.com> >> >> This patch requires your previous "always present" patchset, right? > > On boards where the _STA for INT0002 returns 0, that patchset > is needed for the driver to bind to the device, so yes it has > a functional requirement on some (most it seems) board, but > only at run-time, it can be merged independently. > >> Some nitpicking: >> >>> --- /dev/null >>> +++ b/drivers/platform/x86/intel_int0002.c >> ... >>> +struct int0002_data { >>> + struct spinlock lock; >>> + struct device *dev; >>> + const struct x86_cpu_id *cpu_id; >>> + acpi_handle handle; >>> + char ev_name[5]; >> >> Are 5 bytes enough? I see the code below: >> >>> + snprintf(data->ev_name, sizeof(data->ev_name), "_%c%02X", >>> + res->data.gpio.triggering ? 'E' : 'L', >>> + res->data.gpio.pin_table[0]); >> >> So it counts 6 including NUL. > > The name is e.g. _E02 or _L02 so 5 bytes including the NUL is enough :) > >> >>> +static int int0002_probe(struct platform_device *pdev) >>> +{ >>> + struct acpi_buffer buf = { ACPI_ALLOCATE_BUFFER, NULL }; >>> + struct device *dev = &pdev->dev; >>> + struct int0002_data *data; >>> + struct acpi_resource *res; >>> + acpi_status status; >>> + acpi_handle hdl; >>> + int irq, ret; >>> + >>> + data = devm_kzalloc(dev, sizeof(*data), GFP_KERNEL); >>> + if (!data) { >>> + dev_err(dev, "can't allocate memory for int0002\n"); >> >> The error message is mostly superfluous. > > Agreed, will fix for v2. > >>> +static int int0002_runtime_suspend(struct device *dev) >>> +{ >>> + return 0; >>> +} >>> + >>> +static int int0002_runtime_resume(struct device *dev) >>> +{ >>> + return 0; >>> +} >>> + >>> +static const struct dev_pm_ops int0002_pm_ops = { >>> + .runtime_suspend = int0002_runtime_suspend, >>> + .runtime_resume = int0002_runtime_resume, >>> +}; >> >> Do we need these runtime PM? If not, we can remove the header >> inclusion, too. > > We enable runtime pm for this device, at which point these > are mandatory. Correction, I thought the driver was enabling runtime pm but it does not, so these can simply be removed. Will fix for v2. Regards, Hans -- To unsubscribe from this list: send the line "unsubscribe linux-acpi" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/drivers/platform/x86/Kconfig b/drivers/platform/x86/Kconfig index 53afa78..c400e29 100644 --- a/drivers/platform/x86/Kconfig +++ b/drivers/platform/x86/Kconfig @@ -889,6 +889,16 @@ config INTEL_PMC_CORE Supported features: - SLP_S0_RESIDENCY counter. +config INTEL_INT0002 + tristate "Intel INT0002 Virtual GPIO ACPI device driver" + depends on X86 && ACPI + ---help--- + Some peripherals on Baytrail and Cherrytrail platforms signal + PME to the PMC to wakeup the system. When this happens software + needs to explictly clear the interrupt source to avoid an IRQ + storm on IRQ 9. This is modelled in ACPI through the INT0002 + ACPI device. This driver implements the clearing of the IRQ. + config IBM_RTL tristate "Device driver to enable PRTL support" depends on PCI diff --git a/drivers/platform/x86/Makefile b/drivers/platform/x86/Makefile index 6731893..de4ffb5 100644 --- a/drivers/platform/x86/Makefile +++ b/drivers/platform/x86/Makefile @@ -51,6 +51,7 @@ obj-$(CONFIG_INTEL_SCU_IPC) += intel_scu_ipc.o obj-$(CONFIG_INTEL_SCU_IPC_UTIL) += intel_scu_ipcutil.o obj-$(CONFIG_INTEL_MFLD_THERMAL) += intel_mid_thermal.o obj-$(CONFIG_INTEL_IPS) += intel_ips.o +obj-$(CONFIG_INTEL_INT0002) += intel_int0002.o obj-$(CONFIG_XO1_RFKILL) += xo1-rfkill.o obj-$(CONFIG_XO15_EBOOK) += xo15-ebook.o obj-$(CONFIG_IBM_RTL) += ibm_rtl.o diff --git a/drivers/platform/x86/intel_int0002.c b/drivers/platform/x86/intel_int0002.c new file mode 100644 index 0000000..76220db --- /dev/null +++ b/drivers/platform/x86/intel_int0002.c @@ -0,0 +1,232 @@ +/* + * Intel INT0002 "Virtual GPIO" driver + * + * Copyright (C) 2017 Hans de Goede <hdegoede@redhat.com> + * + * Loosely based on android x86 kernel code which is: + * + * Copyright (c) 2014, Intel Corporation. + * + * Author: Dyut Kumar Sil <dyut.k.sil@intel.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + * Some peripherals on Baytrail and Cherrytrail platforms signal PME to the + * PMC to wakeup the system. When this happens software needs to clear the + * PME_B0_STS bit in the GPE0a_STS register to avoid an IRQ storm on IRQ 9. + * + * This is modelled in ACPI through the INT0002 ACPI device, which is + * called a "Virtual GPIO controller" in ACPI because it defines the event + * handler to call when the PME triggers through _AEI and _L02 / _E02 + * methods as would be done for a real GPIO interrupt. + * + * This driver will bind to the INT0002 device, call the ACPI event handler + * for the wakeup and clear the interrupt source avoiding the irq storm. + */ + +#include <asm/cpu_device_id.h> +#include <asm/intel-family.h> +#include <asm/io.h> +#include <linux/acpi.h> +#include <linux/interrupt.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/platform_device.h> +#include <linux/pm_runtime.h> +#include <linux/slab.h> + +#define GPE0A_PME_STS_BIT 0x2000 +#define GPE0A_PME_EN_BIT 0x2000 +#define GPE0A_STS_PORT 0x420 +#define GPE0A_EN_PORT 0x428 + +#define ICPU(model) { X86_VENDOR_INTEL, 6, model, X86_FEATURE_ANY, } + +static const struct x86_cpu_id int0002_cpu_ids[] = { +/* + * Limit ourselves to Cherry Trail for now, until testing shows we + * need to handle the INT0002 device on Baytrail too. + * ICPU(INTEL_FAM6_ATOM_SILVERMONT1), * Valleyview, Bay Trail * + */ + ICPU(INTEL_FAM6_ATOM_AIRMONT), /* Braswell, Cherry Trail */ + {} +}; + +struct int0002_data { + struct spinlock lock; + struct device *dev; + const struct x86_cpu_id *cpu_id; + acpi_handle handle; + char ev_name[5]; +}; + +static void int0002_irq_enable(struct int0002_data *data, bool enable) +{ + unsigned long flags; + u32 gpe_en_reg; + + spin_lock_irqsave(&data->lock, flags); + + gpe_en_reg = inl(GPE0A_EN_PORT); + if (enable) + gpe_en_reg |= GPE0A_PME_EN_BIT; + else + gpe_en_reg &= ~GPE0A_PME_EN_BIT; + outl(gpe_en_reg, GPE0A_EN_PORT); + + spin_unlock_irqrestore(&data->lock, flags); +} + +static irqreturn_t int0002_irq_handler(int irq, void *handler_data) +{ + struct int0002_data *data = handler_data; + u32 gpe_sts_reg; + + gpe_sts_reg = inl(GPE0A_STS_PORT); + if (!(gpe_sts_reg & GPE0A_PME_STS_BIT)) + return IRQ_NONE; + + int0002_irq_enable(data, false); + + return IRQ_WAKE_THREAD; +} + +static irqreturn_t int0002_irq_thread(int irq, void *handler_data) +{ + struct int0002_data *data = handler_data; + acpi_status status; + + /* Don't call ACPI event handler on Baytrail? Taken from Android-x86 */ + if (data->cpu_id->model != INTEL_FAM6_ATOM_SILVERMONT1) { + status = acpi_evaluate_object(data->handle, data->ev_name, + NULL, NULL); + if (ACPI_FAILURE(status)) + dev_err(data->dev, "Error calling %s\n", data->ev_name); + } + + /* Ack and then re-enable IRQ */ + outl(GPE0A_PME_STS_BIT, GPE0A_STS_PORT); + int0002_irq_enable(data, true); + + return IRQ_HANDLED; +} + +static int int0002_probe(struct platform_device *pdev) +{ + struct acpi_buffer buf = { ACPI_ALLOCATE_BUFFER, NULL }; + struct device *dev = &pdev->dev; + struct int0002_data *data; + struct acpi_resource *res; + acpi_status status; + acpi_handle hdl; + int irq, ret; + + data = devm_kzalloc(dev, sizeof(*data), GFP_KERNEL); + if (!data) { + dev_err(dev, "can't allocate memory for int0002\n"); + return -ENOMEM; + } + + spin_lock_init(&data->lock); + data->dev = dev; + + /* Menlow has a different INT0002 device? <sigh> */ + data->cpu_id = x86_match_cpu(int0002_cpu_ids); + if (!data->cpu_id) + return -ENODEV; + + data->handle = ACPI_HANDLE(dev); + if (!data->handle) + return -ENODEV; + + irq = platform_get_irq(pdev, 0); + if (irq < 0) { + dev_err(dev, "Error getting IRQ: %d\n", irq); + return irq; + } + + status = acpi_get_event_resources(data->handle, &buf); + if (ACPI_FAILURE(status)) { + dev_err(dev, "Error getting acpi event resources\n"); + return -ENODEV; + } + + /* Find the "GPIO interrupt" event handler to call upon PME */ + ret = -ENODEV; + for (res = buf.pointer; + res && (res->type != ACPI_RESOURCE_TYPE_END_TAG); + res = ACPI_NEXT_RESOURCE(res)) { + + if (res->type != ACPI_RESOURCE_TYPE_GPIO || + res->data.gpio.connection_type != + ACPI_RESOURCE_GPIO_TYPE_INT) + continue; + + snprintf(data->ev_name, sizeof(data->ev_name), "_%c%02X", + res->data.gpio.triggering ? 'E' : 'L', + res->data.gpio.pin_table[0]); + + status = acpi_get_handle(data->handle, data->ev_name, &hdl); + if (ACPI_SUCCESS(status)) { + ret = 0; + break; + } + } + + ACPI_FREE(buf.pointer); + + if (ret) { + dev_err(dev, "Error could not find event handler\n"); + return ret; + } + + ret = devm_request_threaded_irq(dev, irq, + int0002_irq_handler, int0002_irq_thread, + IRQF_SHARED, "INT0002", data); + if (ret) { + dev_err(dev, "Error requesting IRQ %d: %d\n", irq, ret); + return ret; + } + + int0002_irq_enable(data, true); + + return 0; +} + +static int int0002_runtime_suspend(struct device *dev) +{ + return 0; +} + +static int int0002_runtime_resume(struct device *dev) +{ + return 0; +} + +static const struct dev_pm_ops int0002_pm_ops = { + .runtime_suspend = int0002_runtime_suspend, + .runtime_resume = int0002_runtime_resume, +}; + +static const struct acpi_device_id int0002_acpi_ids[] = { + { "INT0002", 0 }, + { }, +}; +MODULE_DEVICE_TABLE(acpi, int0002_acpi_ids); + +static struct platform_driver int0002_driver = { + .driver = { + .name = "Intel INT0002 Virtual GPIO", + .pm = &int0002_pm_ops, + .acpi_match_table = ACPI_PTR(int0002_acpi_ids), + }, + .probe = int0002_probe, +}; + +module_platform_driver(int0002_driver); + +MODULE_AUTHOR("Hans de Goede <hdegoede@redhat.com>"); +MODULE_DESCRIPTION("Intel INT0002 Virtual GPIO driver"); +MODULE_LICENSE("GPL");
Some peripherals on Baytrail and Cherrytrail platforms signal PME to the PMC to wakeup the system. When this happens software needs to clear the PME_B0_STS bit in the GPE0a_STS register to avoid an IRQ storm on IRQ 9. This is modelled in ACPI through the INT0002 ACPI device. This commit adds a driver which will bind to that device, call the ACPI event handler for the wakeup and clear the interrupt source avoiding the irq storm. Signed-off-by: Hans de Goede <hdegoede@redhat.com> --- drivers/platform/x86/Kconfig | 10 ++ drivers/platform/x86/Makefile | 1 + drivers/platform/x86/intel_int0002.c | 232 +++++++++++++++++++++++++++++++++++ 3 files changed, 243 insertions(+) create mode 100644 drivers/platform/x86/intel_int0002.c